Published by Oxford University Press on behalf of the International Epidemiological Association
Ó The Author 2006; all rights reserved. Advance Access publication 9 May 2006
International Journal of Epidemiology 2006;35:1059–1065
doi:10.1093/ije/dyl091
Epidemiology of a primary pneumonic plague
in Kantoshu, Manchuria, from 1910 to 1911:
statistical analysis of individual records
collected by the Japanese Empire
Hiroshi Nishiura1–3
Accepted
5 April 2006
Background Among the potential uses of Yersinia pestis, intentional release of its aerosolized
form, causing person-to-person transmission, is thought to be the most
threatening. With the current rarity of pneumonic plague epidemics, our
epidemiological knowledge remains insufficient for detailed characterization of
effective control measures.
Methods
Temporal patterns and key biological parameters of a pneumonic plague
epidemic in Manchuria from 1910 to 1911 were analysed based on historical
records collected by Kanto Totokufu, the administration of the Japanese Empire
in Manchuria at that time. The serial intervals were fitted to gamma distribution
using the maximum likelihood method, and time-delay distributions from
onset-to-admission, admission-to-death, and onset-to-death were investigated.
Results
Whereas a total of 228 cases were diagnosed with pneumonic plague in areas
under direct control of the Japanese Empire, 4781 cases were also recorded in
surrounding areas. Although the epidemic grew exponentially in the early
phase, the average doubling time steadily increased reflecting successful control
efforts. The estimated mean serial interval (and standard deviation) was 5.7 (3.6)
days. All cases with known dates of onset were admitted to hospital within 4 days
after onset, and the mean time from onset to admission was 1.1 (0.4) days.
Conclusions The increase in doubling time demonstrates the efficient and rapid countermeasures employed. Since the short interval from onset to death implies the
importance of rapid responses, the challenge in confronting a future bioterrorist
attack is to implement rapid and appropriate integration of control measures
both at the individual and community level to prevent further transmissions as
well as lower case fatality.
Keywords
Epidemiology, pneumonic plague, epidemics, statistical distributions, confidence
intervals, historical events
Although the annual incidence of plague has gradually
decreased,1 there is a plausible worldwide threat of bioterrorist
1
2
3
Department
of
Medical
Biometry,
University
Westbahnhofstrasse 55, Tübingen, D-72070, Germany.
of
Tübingen,
Graduate School of Health Sciences, Hiroshima University, Kasumi 1-2-3,
Minamiku, Hiroshima, 734-8551, Japan.
Research Center for Tropical Infectious Diseases, Nagasaki University
Institute of Tropical Medicine, Sakamoto 1-12-4, Nagasaki, 852-8523,
Japan. E-mail: [email protected]
attack using the causative agent Yersinia pestis.2,3 The Centers
for Disease Control and Prevention (CDC) classified this agent
as category A owing to its biological as well as epidemiological
characteristics and potential impact.4 Among its potential uses,
intentional release of aerosolized Y. pestis, causing personto-person transmission of primary pneumonic plague, is
thought to be the most threatening. The proportion of pneumonic plague-associated case fatalities without appropriate
treatment within 24 h after onset is thought to be ~100%.5
There are still no specific methods of primary prevention such
1059
1060
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
(a)
(b)
Figure 1 Political map of Far East Asia from 1910 to 1911. (a) Political areas in Far East Asia from 1910 to 1911. Both Japanese and Russian
armies evacuated Manchuria after the Treaty of Portsmouth in 1905, but the Japanese Empire remained in Liaodong Peninsula, which contains
Dalian and Port Arthur, under control of the headquarters of the Empire, Kantoshu (small boxed area). The power of Kantoshu progressed
northward (large boxed area) owing to penetration of the Japanese Empire with the building of the Manchurian Railway; this became the
preceding political zone of Manchukuo after 1931. After falling under the Japanese Protectorate between 1905 and 1910, Korea was annexed as a
part of the Japanese Empire from 1910 to 1945. (b) Enlarged map of Kantoshu and areas illegally penetrated up to 1910–11. The Manchurian
Railway connected Changchun and Dalian. Only Liaodong Peninsula was completely under the direct control of Kantoshu. Other northern cities,
including Mukden, Fushun, Changchun, and Gongzhuling, were partly occupied by the Japanese Empire (around train stations only);
30
surrounding areas remained under political control of the Chinese Empire. The map was drawn by the author with reference to a historical map
as vaccine.6 and the appearance of multi-drug-resistant strains
implies potential difficulties of treatment in the case of future
attacks.7 Since real-time observations of large-scale primary
pneumonic plague epidemics are now extremely rare, historical
records can assist in determining key information. Reviewing
historical epidemic records has been effective in the search for
potential control strategies,8,9 and accumulation of different
epidemic records has helped obtain estimates of important
biological and epidemiological parameters.10,11 Nevertheless,
owing to the scarcity of reports and relatively small size of
each epidemic to date, our knowledge on the epidemiology of
pneumonic plague is still insufficient for detailed characterization of effective control measures, and, hence, further
investigation is required. To date, the biggest epidemic of
pneumonic plague to be documented with bacteriological
confirmation was that in Manchuria from 1910 to 1911.12
Whereas the statistical records of this epidemic in the report of
the International Plague Conference held in Mukden,
Manchuria, 1911, were fragmentarily collected from different
locations (i.e. from several plague hospitals) owing to the large
number of deaths and rapidity of spread beyond the capacity of
the health authorities,13 detailed documentation was also
performed within the political zone of the Japanese Empire in
Manchuria at that time. In the present study, total and
individual case records are analysed based on the report of
Kanto Totokufu,14 the administration section in charge of
the political areas of the Japanese Empire in Manchuria, with
particular emphasis on key distributions and parameter
estimations over time.
Materials and methods
Brief history of Kantoshu and Kanto Totokufu
After the end of the Russo-Japanese War with the signing of
the Treaty of Portsmouth in 1905, several parts of Manchuria,
such as Dalian, came under direct political control of Kantoshu,
an annex of the Japanese Empire.15 Although Kantoshu
represented the political area of the Japanese Empire in
Liaodong Peninsula, its overall power progressed northward
through construction of the South Manchurian Railway and
subsequent penetration of the Empire (Figure 1a). In this
enlarged political area (i.e. the headquarters of Kantoshu in
Liaodong Peninsula and the areas around train stations in
several cities in Manchuria; see Figure 1b), Kanto Totokufu,
which was established in 1906, was placed in charge of
controlling administration, including public health practices.
Remaining areas were under the political control of the
Chinese Empire. At the time of the pneumonic plague
epidemic in Manchuria, Kanto Totokufu organized a temporary
quarantine section. Task force members were also given
the responsibility of performing investigations and summarizing the statistical records of the epidemic in and around
areas under direct control of the Japanese Empire.14
Data source: description of the historical
observation
In 1910, owing to the uncontrollability of the plague
epidemic in Chinese controlled areas, Kanto Totokufu urgently
requested medical professionals and their assistants from
the Japanese Islands. Organizations receiving these requests
included the Japanese Ministry of the Interior, the National
Institute of Communicable Diseases (presently The Institute of
Medical Science, The University of Tokyo), and Osaka and
Hyogo prefectures. Since Kanto Totokufu viewed the
epidemic as a serious political and economic threat,16 an
extraordinary number of professionals were assigned to assist in
the disease control tasks. A total of 2182 police guards
were engaged in contact tracing and quarantine, while
7339 others worked for the temporary quarantine section.
STATISTICAL ANALYSIS OF INDIVIDUAL RECORDS COLLECTED BY THE JAPANESE EMPIRE
The total number of documented pneumonic plague cases was
5009, including those from areas surrounding Kantoshu. Based
on efforts within areas under direct political control of
Kantoshu,14 individual data were obtained for 228 cases,
enabling detailed epidemiological investigations. Individual
information included name, sex, age, diagnostic methods,
source of infection, and dates of onset, discovery, hospital
admission, and death. The 228 individual records covered
observations made after notification of the index case on
December 31, 1910 until the end of the epidemic on April 17,
1911. The remaining 4781 cases were diagnosed in surrounding
areas and provided the date of notification only. Since the
epidemic became remarkable in other areas of Manchuria from
early October 1910,17 public health measures such as contact
tracing, quarantine, and isolation were extensively performed
at the time of record collection. At this time, there was no
specific effective treatment for pneumonic plague.18
Statistical analysis
As descriptive information, temporal distributions of pneumonic plague cases in and around areas under direct control of
Kantoshu were obtained. Spatial information was obtained
only within directly controlled areas, and reliable demographic
data were unavailable. Based on the total daily number of
reported cases, doubling time, T, was given by:
tk t0
ð1Þ
T¼
log2 NN0k
where Nk and N0 are the cumulative number of cases at times tk
and t0, respectively. Since 88 of the 228 records included a
definitive description of the potential source of infection based
on room sharing in a hospital or at home, the serial interval,
the time interval between infection of one person and infection
of others by this individual (or the time from symptom onset in
the index case to symptom onset in a secondary case),19 was
estimated from these data. The interval was fitted to a gamma
distribution using the maximum likelihood method. The
Cramer–Smirnov–von Mises test was employed to assess the
goodness-of-fit.20 Time-delay distributions (from onsetto-admission, admission-to-death, and onset-to-death) were
also investigated. Since these distributions did not reasonably
fit standard statistical distributions, estimates were obtained as
sample means and standard deviation (SD), not assuming
specific distributions. To estimate the parameters of the above
three time-delay distributions, only those cases for which required
information was available were used, ignoring incomplete
observations. The statistical data were analysed using the statistical
software JMP IN version 5.1 (SAS Institute Inc., Cary, NC).
Results
Descriptive data and temporal distributions
Among a total of 5009 recorded plague cases, 228
were diagnosed as having pneumonic plague in areas
under direct control of Kantoshu. Figures 2a and b show the
temporal distributions in surrounding and directly controlled
areas, respectively. These results are based on the dates of
notification, and, hence, include delays in discovery of
death, diagnosis, and reporting. Notifications peaked in late
1061
January 1911 in directly controlled areas and in mid-February
1911 in surrounding areas, while the records of other Chinese
controlled
areas
peaked
in
late
November
and
December, 1910.13 The cumulative number of cases in both
areas on a logarithmic scale is given in Figure 2c. The epidemic
showed an exponential increase in the early phase. The average
doubling time for the total number of reported cases is given in
Figure 2d. After a drop in early January, the doubling time
continued to increase over time. Supplementary data 1 shows
the number of cases with time used to create these temporal
distributions (Supplementary data available at IJE online).
Detailed individual records were obtained for all 228 cases
diagnosed in the directly controlled areas. The mean age and
SD of these cases were 33.2 and 10.6 years, respectively; a crude
age distribution is shown in Figure 2e. The majority of cases
consisted of adults aged between 21 and 50 years (n 5 200,
87.7%). Only two cases were female (0.9%). In directly
controlled areas (Figure 1b), the majority of cases were diagnosed
in Changchun (n 5 105, 46.1%) and Dalian (n 5 65, 28.5%);
other places included Fushun (n 5 18, 7.9%), Mukden (presently
Shenyang; n 5 13, 5.7%), and Gongzhuling (n 5 12, 5.3%).
Diagnosis was made mainly by sputum examination under a
microscope or sputum culture of Y. pestis (Table 1). Numbers
diagnosed using each method in Table 1 denote successful
confirmation with shown diagnostic tests. Nearly 40% of cases
were confirmed by a combination of these two methods.
However, diagnoses of 51 cases (23.1%) were confirmed by
pathological examination after death, and, of these, 66.7% were
further confirmed using bacterial culture of samples. Only 16
cases (7.2%) were diagnosed by means of clinical examinations
alone. The records suggest that clinical diagnoses were almost
all based on signs of respiratory failure with bloody sputum.
Distributions of the serial interval and time delays
The maximum likelihood estimates of the mean serial interval
and SD were 5.7 [95% confidence interval (95% CI) 4.9–6.4]
and 3.6 days, respectively. Therefore, the shape and scale
parameters of the gamma distribution were 2.5 (95% CI
1.9–3.3) and 2.2 (95% CI 1.7–3.1), respectively. As a visual
comparison, the Kaplan–Meier survival curve together with
the parametric cumulative survival probability is given in
Figure 3a. The mean [and corresponding standard error (SE)]
in the non-parametric method was also estimated as 5.7 (0.4)
days. Cramer–von Mises goodness-of-fit test revealed no
significant deviation between the observed and expected
distributions (W2 5 0.17, P 5 0.25). Supplementary data 2
includes the serial intervals for the 88 investigated cases
(Supplementary data available at IJE online).
Figures 3b–d show the frequencies of time from onsetto-death, onset-to-admission, and admission-to-death on a
discrete time scale. The estimated mean and SD for these
three time periods were: 2.3 and 1.7 days; 1.1 and 0.4 days; and
0.9 and 1.1 days, respectively. In each case, there was no
significant association with age (P 5 0.55, 0.13, and 0.77,
respectively). The main reasons for exclusion from Figures 3b–d
were cases with unknown dates of onset (n 5 62, 27.2%), those
not admitted to hospital and who died soon or suddenly after
discovery (n 5 85, 37.3%), and those found dead (n 5 4, 1.8%).
All confirmed cases investigated here resulted in death.
1062
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
(a)
(b)
300
20
250
Number of cases
Number of cases
15
200
150
100
10
5
50
0
30
/1
2
06
/0
1
13
/0
1
20
/0
1
27
/0
1
03
/0
2
10
/0
2
17
/0
2
24
/0
2
03
/0
3
10
/0
3
17
/0
3
24
/0
3
31
/0
3
07
/0
4
14
/0
4
30
/1
2
06
/0
1
13
/0
1
20
/0
1
27
/0
1
03
/0
2
10
/0
2
17
/0
2
24
/0
2
03
/0
3
10
/0
3
17
/0
3
24
/0
3
31
/0
3
07
/0
4
14
/0
4
0
Date of notif ication
(c)
10
Doubling time
10000
100
1
Date of notif ication
(e)
/0
4
15
/0
4
/0
4
01
08
/0
3
/0
3
25
/0
3
11
18
/0
2
/0
3
04
25
/0
2
/0
2
11
18
/0
1
/0
2
04
/0
1
21
28
/0
1
/0
1
07
/1
2
31
30
/1
06 2
/0
13 1
/0
20 1
/0
27 1
/0
03 1
/0
10 2
/0
17 2
/0
24 2
/0
03 2
/0
10 3
/0
17 3
/0
24 3
/0
31 3
/0
07 3
/0
14 4
/0
4
1
14
Cumulative number of cases
Date of notification
(d)
Date
100
Number of cases
80
60
40
20
ov
er
70
61
-7
0
51
-6
0
41
-5
0
31
-4
0
21
-3
0
11
-2
0
5
610
1-
un
de
r
1
0
Age [years]
Figure 2 Epidemiological description of the primary pneumonic plague epidemic in and around Kantoshu, Manchuria, from 1910 to 1911.
(a) Temporal distribution of plague notifications in surrounding areas (n 5 4781) and (b) areas under direct control of Kantoshu (n 5 228).
(c) Logarithmic temporal distribution of the cumulative number of cases in and around the directly controlled areas (n 5 5009). (d) Doubling
time of the primary pneumonic plague epidemic in and around the directly controlled areas with time (n 5 5009). (e) Age distribution
of cases in directly controlled areas (n 5 228)
Discussion
Epidemic records of a pneumonic plague epidemic in Manchuria from 1910 to 1911 were analysed based on 5009 cases
included in the report by Kanto Totokufu, the administration in
charge of the political areas of the Japanese Empire in
Manchuria at the time. Several key statistical distributions
were obtained for 228 individual pneumonic plague cases
diagnosed in areas under direct control. All investigated cases
resulted in death. As in recent rigorous studies on severe acute
respiratory syndrome (SARS),21,22 obtained distributions statistically reflect the biological and epidemiological characteristics of the course of a disease and epidemic. For example, in the
examined epidemic, individuals, excluding those found after or
close to death, were efficiently admitted to hospitals within 4 days
STATISTICAL ANALYSIS OF INDIVIDUAL RECORDS COLLECTED BY THE JAPANESE EMPIRE
after onset. As has been discussed previously,23 this reflects the
severity of pneumonic plague, which leads to death within a short
time period without treatment, or extent of the control measures.
The epidemic period investigated in this study saw extensive
interventions implemented almost from when the collection of
official records was started.14 Since there was no specific
treatment for pneumonic plague at this time, most of the efforts
by Kanto Totokufu were devoted to prevention both at the
individual and community level. At the individual level, the
use of a gauze mask was recommended almost as obligatory
practice not only for those participating in public health and
medical practices (e.g. sanitary officers and hospital assistants)
but also the general population.24 Through household rounds,
the notifications of infected and deceased persons were strictly
supervised, and at the same time, an attempt was made to
educate the public about the mode of transmission, disinfection
methods, and rodent traps (partly based on knowledge of
bubonic plague control).14 At the population level, it is
documented that public health interventions included not
only quarantine of potentially infected cases and isolation
measures but also screening and identification of possible
contacts in order to find infected individuals, i.e. contact
tracing.15,25,26 Because the number of hospital beds appeared
to be too few to admit all cases, Kanto Totokufu constructed
temporary camps for isolation purposes only, whereby contact
was kept under strict observation.14 Since pneumonic plague is
known to have spread spatially as a result of construction of the
Manchurian Railway,27 clinical examinations and quarantines
Table 1 Diagnostic methods used to confirm diagnoses of primary
pneumonic plague in Kantoshu, Manchuria, from 1910 to 1911
(n 5 221)
Diagnostic method
n
Percentage
Sputum culture
161
72.9
Microscopic examination of sputum
110
49.8
51
23.1
Blood culture
46
20.8
Clinical diagnosis only
16
7.2
86
38.9
34
15.4
Pathological diagnosis
a
Microscopy and culture of sputum
Anatomy and culture of samples
b
b
1063
a
Pathological diagnosis was made by autopsy and, in part, histopathological
examination.
b
Combined methods.
(b)
(a)
1
0.6
Non-parametric
0.5
Maximum likelihood
(gamma distribution)
0.4
0.6
Frequency
1 - cumulative density
0.8
0.4
0.3
0.2
0.2
0.1
0
0
0
2
4
(c)
6
8
10
12
Serial interval [days]
14
0
16
1
2
3
4
5
6
7
8
9
10
11
12
10
11
12
Time from onset to death [days]
(d)
1
0.5
0.4
0.6
Frequency
Frequency
0.8
0.4
0.2
0.3
0.2
0.1
0
0
1
2
3
4
5
6
7
8
9
Time f rom onset to admission [days]
10
11
12
0
1
2
3
4
5
6
7
8
9
Time from admission to death [days]
Figure 3 Frequency distributions of key epidemiological determinants of primary pneumonic plague based on the records of Kanto
Totokufu, Manchuria, from 1910 to 1911. (a) Non-parametric (broken line) and maximum likelihood gamma probabilities (solid line) of the
serial interval given as the cumulative survival probability (n 5 88). (b) Onset-to-death distribution (n 5 166). (c) Onset-to-admission
distribution (n 5 126). (d) Admission-to-death distribution (n 5 143)
1064
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
were also performed in areas surrounding each train station on
the line.28 Moreover, traffic restrictions were implemented to
limit movement of infectious individuals into highly populated
areas; in particular, complete road blockage was performed in
rural locations where no trained sanitary officers were
placed.14,24 As a result, whereas an exponential increase was
observed in the early stages of the epidemic, the doubling time
slowly but steadily increased over time due to lowered transmission
rates and the effectiveness of the control efforts.29 This increase
reflects eventual control of the epidemic in investigated areas.
The important implications of the present findings are the
requirement for rapid responses during potential future
epidemics. Although the investigated epidemic, which lasted
for more than 3 months, differs from those at present where
effective public health interventions include quick release of
information through mass media, the historical records suggest
the potential for disaster even with considerable preventive
measures both at the individual and community level. Whereas
the investigated epidemic eventually declined to extinction as a
result of concerted efforts, one of the biggest current concerns
over pneumonic plague is the extremely high risk of death
following the short symptomatic period. Even though there are
now several specific treatment methods, even a slight time lag
from onset to admission could result in death. Therefore, the
challenge in confronting a future bioterrorist attack is the
implementation of extremely rapid and appropriate integration
of individual and community interventions. Although primary
pneumonic plague is highly fatal and its aetiological agent
Y. pestis is readily available, epidemiological studies of this
disease as well as investigations of intervention strategies
remain rather limited compared with those for smallpox. In this
context, biological and epidemiological research on pneumonic
plague should be strengthened in the future.
Acknowledgements
I would like to acknowledge Profs Shinichi Yoshida, Wataru
Iijima, Masaaki Shimada, and Masayuki Kakehashi, Ms
Mutsuko Moriwaki, and Dr Shigeki Hanafusa for their support
in data collection. This study was supported by the Banyu
Fellowship Program, which is sponsored by the Banyu Life
Science Foundation International.
KEY MESSAGES
Whereas an exponential increase was observed in the early stages, the doubling time steadily increased over
time reflecting effective control measures.
Pneumonic plague cases, excluding those found after or close to death, were admitted to hospitals within 4 days
after onset.
The challenge in confronting a future bioterrorist attack is rapid and appropriate integration of control measures
both at the individual and community level to prevent further transmissions as well as lower case fatality.
References
1
2
3
4
5
6
7
8
9
World Health Organization. Human plague in 2002 and 2003. Wkly
Epidemiol Rec 2004;79:301–06.
Inglesby TV, Dennis DT, Henderson DA et al. Plague as a biological
weapon: Medical and public health management. Working Group on
Civilian Biodefense. JAMA 2000;283:2281–90.
McGovern TW, Friedlander AM. Chapter 23. Plague. In: Siedell FR,
Takafuji ET, Franz DR (eds). Medical Aspects of Chemical and Biological
Warfare. Washington DC: Walter Reed Army Medical Center, 1997,
pp. 479–502.
Centers for Disease Control and Prevention. Biological and chemical
terrorism: strategic plan for preparedness and response. Recommendations of the CDC Strategic Planning Workgroup. MMWR Recomm Rep
2000;49:RR–4.
Pollitzer R. Plague. Geneva: World Health Organization, 1954.
10
Gani R, Leach S. Epidemiologic determinants for modeling
pneumonic plague outbreaks. Emerg Infect Dis 2004;10:608–14.
11
Nishiura H, Schwehm M, Kakehashi M, Eichner M. Transmission
potential of primary pneumonic plague: time-inhomogeneous evaluation based on historical documents of the transmission network.
J Epidemiol Community Health 2006;60:doi:10.1136/jech.2005.042424.
12
Wu L-T. A Treatise on Pneumonic Plague. Geneva: League of Nations
Health Organization, 1926.
13
Petrie GF. An epidemiological review of the epidemic of pneumonic
plague in northern China, 1910 to 1911. PART III. Summary of
knowledge gained from the study of the epidemic. Chapter I. In:
Report of the international plague conference. Held at Mukden, April, 1911.
Manila: Bureau of Printing, 1912, pp. 407–27.
14
Temporary Quarantine Section, Kanto Totokufu. Epidemic Record of
Plague During 1910–1911 (Meiji 43,4-nen ‘Pest’ Ryuko-shi). Dalian:
Manchurian Daily Press (Manshu Nichi-nichi Shinbunsya), 1912
(in Japanese).
Titball RW, Williamson ED. Vaccination against bubonic and
pneumonic plague. Vaccine 2001;19:4175–84.
15
Galimand M, Guiyoule A, Gerbaud G et al. Multidrug resistance in
Yersinia pestis mediated by a transferable plasmid. N Engl J Med
1997;337:677–80.
16
Kool JL. Risk of person-to-person transmission of pneumonic plague.
Clin Infect Dis 2005;40:1166–72.
17
Levinson ME. Lessons learned from history on mode of transmission for control of pneumonic plague. Curr Infect Dis Rep
2000;2:269–71.
Nathan CF. Plague Prevention and Politics in Manchuria 1910–31.
Cambridge: East Asian Research Center, Harvard University, 1967.
Iijima W. Chapter 5. Epidemic of pneumonic plague and politicized
public health: from 1910 to 1911. In: Plague and Modern China (Pest to
Kindai Chugoku). Tokyo: Kenbun Press, 2000 (in Japanese).
Kasai K. The epidemiology of the plague in Southern Manchuria.
PART I. Presentation of evidence regarding the epidemic. Session III.
In: Report of the International Plague Conference. Held at Mukden, April,
1911. Manila: Bureau of Printing, 1912, pp. 33–35.
STATISTICAL ANALYSIS OF INDIVIDUAL RECORDS COLLECTED BY THE JAPANESE EMPIRE
18
19
20
21
22
23
Christie D. Clinical data. PART I. Presentation of evidence regarding
the epidemic. Session IX. In: Report of the International Plague
Conference. Held at Mukden, April, 1911. Manila: Bureau of Printing,
1912, pp. 163–66.
Anderson RM, May RM. Infectious Diseases of Humans: Dynamics and
Control. Oxford: Oxford University Press, 1991.
Stephens MA. Use of the Kolomogorov-Smirnov, Cramer-von Mises
and related statistics without extensive tables. J R Stat Soc Ser B
1970;32:115–22.
24
25
26
27
Donnelly CA, Ghani AC, Leung GM et al. Epidemiological determinants of spread of causal agent of severe acute respiratory syndrome
in Hong Kong. Lancet 2003;361:1761–66.
Leung GM, Hedley AJ, Ho LM et al. The epidemiology of
severe acute respiratory syndrome in the 2003 Hong Kong
epidemic: An analysis of all 1755 patients. Ann Intern Med
2004;141:662–73.
Wu L-T. Problems of pneumonic plague. In: Cunningham LCJ (ed.).
Transactions of the Seventh Congress Held in British India, December, 1927.
Far Eastern Association of Tropical Medicine. Volume II. Calcutta:
Thacker’s Press & Directories, 1928, pp. 22–34.
28
29
30
1065
Wu L-T, Chun JWH, Pollitzer R. Clinical observations upon the
Manchurian plague epidemic, 1920–21. J Hyg 1923;21:289–306.
Wu L-T. Manchurian Plague Prevention Service: Memorial volume
1912–1932. Shanghai: National Quarantine Service, 1934.
Wu L-T, Chun JWH, Pollitzer R, Wu CY. Plague: A manual for Medical
and Public Health Workers. Shanghai: Weishengshu National Quarantine Service, Mercury Press, 1936.
Stanley A. Quarantine measures at Shanghai against northern ports
infected with pneumonic plague. PART I. Presentation of evidence
regarding the epidemic. Session XV. In: Report of the International
Plague Conference. Held at Mukden, April, 1911. Manila: Bureau of
Printing, 1912, pp. 305–21.
Wu L-T. Plague in the Orient with special reference to the
Manchurian outbreaks. J Hyg 1922;26:62–76.
Galvani AP, Lei X, Jewell NP. Severe acute respiratory syndrome:
Temporal stability and geographic variation in case-fatality rates and
doubling times. Emerg Infect Dis 2003;9:991–94.
Ward AW, Proth GW, Leathes S, Benians EA. The Cambridge
Modern History Atlas. Cambridge: Cambridge University Press, 1912.
Copy of poster published by the Ministry of Interior, Japanese Empire, from the influenza epidemic in 1918.
The top figure promotes the use of masks inside the train. Because the use of masks has been known to be effective to prevent droplet infections
during the Manchurian plague outbreak, and because the train has been recognized as the most dangerous setting through the experience of
Manchurian plague outbreak, the government suggested all individuals should wear the mask during the influenza epidemic. The bottom figure
promotes gargling for those returning from outside. Whereas the effectiveness of gargling is still questionable even at present, the government
encouraged everyone to wash their tonsils everytime after walking outside. The Japanese words in the middle are almost equivalent to the slogan.
This says ‘Wear the mask within the train and crowds!’ and ‘Don’t forget gargling after coming back from outside!’. The title at the bottom is ‘Mask
and gargling’.